All case studies
Illustrative scenario

An AI answer needs a responsible reviewer

This governance scenario gives approved AI use a defined task, permitted information and a review rule before a draft is used in professional work.

  • Adoption & control
  • AI & knowledge

Designed change

A practical AI-use guide linking each approved task to its information rules, required checks and responsible reviewer.

Illustrative scenario combining common practice needs. Software shown is a suggested stack.

Write the rule beside the task

A broad instruction to use AI carefully leaves too much to interpretation. This design would name the tasks staff could try, the information permitted in the tool and the uses requiring approval. Practice examples would show the difference between a working draft, a source-supported conclusion and an output that needs to be rejected.

Make review part of ordinary work

Before using an output, the responsible person would check it against the relevant sources and record material corrections. Uncertain answers and unexpected disclosures would follow an agreed escalation route. The guide would change as testing revealed limits, with the chosen tool's actual settings and terms reviewed separately from the workflow illustration.

The practical lesson

Give each approved AI task a review rule. The person using the output still needs to understand and stand behind it.

Recognise the problem?

Talk through the process, the constraints and what a useful change would look like.

Discuss your practice